Hear me out: Respondents are dead. Long live MR robots
Ever had an idea that you know is genius, but everybody else thinks is crazy? Here is your chance to share it with the world of research. This month Will Goodhand of BrainJuicer proposes building an army of robo-respondents.
What’s the big idea?
Simple. No questions asked, no respondents required, just lovingly created research robots constantly searching the web for people like them, gathering their thoughts, blogs, photos, music and videos and adopting them as the robot’s own to produce a living, breathing single-person synthesis of the many thousands of people who fit the profile.
Yes. It’s a way of taking the world’s single richest natural research resource – the web – and filtering the teeming mass of humanity to produce insights, ideas and understanding.
“The brilliance of DigiViduals is that they don’t just observe one person but all the people on the internet. They crowdsource their own personality”
Sounds like sci-fi. How does it work?
These research robots, or ‘DigiViduals’ as we call them, are Twitter characters constructed from keywords and emotions representing a particular type of person. Every hour or so we take the temperature of the target audience to establish which of its keywords our character should currently be ‘feeling’. Then we send them out to look for people tweeting using those words. The character finds a tweet matching its profile, adopts it as its own and re-tweets it. It even finds associated media on the internet – pictures, videos, references to venues near the character’s location – to bring the target audience to life.
Doesn’t that result in a lot of meaningless noise?
The DigiViduals are designed to filter signal from noise, and create one believable, easily understandable person from the sea of humanity on the web. They bring back the social objects that are most relevant to their character, but researchers’ skills are still crucial to make sense of that material.
Aren’t you just creating a poor man’s ethnographer?
The brilliance of DigiViduals is that they don’t just observe one person but all the people on the internet. They crowdsource their own personality. Ultimately, it’s all human-generated content and, unlike classic research, you often get things that respondents wouldn’t reveal to you straight out: their attitudes to sex, implicit associations and so on.
I see. You want to use your robot slaves to spy on people.
Not at all. Why are all these consumers posting, blogging and tweeting? Because the internet has enabled them to express themselves more fully. They want to be listened to. And at last somebody is listening, and better understanding them and the products and services they might want as a result. It’s just what market research has always done, only now it can be done more quickly, more efficiently and on a previously unimaginable scale - all without boring respondents to death with another survey.
Aren’t there already tools I can use to see what’s being said on the net?
Current approaches track and sift to find keywords (such as brand names) but DigiViduals have the ability to target specific demographic groups and identify context - meaning they go deeper, gaining insights into the attitudinal and emotional territories of target groups. Without ever asking a single question, DigiViduals gather and make sense of the enormous amount of unprompted conversations and dialogue online. You can use them for one-off projects or leave them to gather data continuously, creating a virtual insight tracker. And researchers and brand managers can follow them, one for each of their segments.
But what if the robots rebel and enslave humanity?
There is that risk. But in the meantime imagine a marketing department where smartphones carrying the latest Twitter feed from each DigiVidual are attached to robots equipped with text-to-speech software. It’s impossible not to listen to the consumer when they’re looking over your shoulder and whispering in your ear.
Share your vision with us: firstname.lastname@example.org